Can we agree headphone amplifiers are solved?

[solderdude]

@solderdude I would be interested in hearing real world measurements of different headphones. Like measuring power levels and then measure real headphone output for the given power levels and see where distortion and compression sets in.

My measurement setup is not suited for this. It is designed to measure at 90dB SPL and clips at 105dB.
When you reach clipping levels of amps you either are using wimpy amps or you listen at abnormally high levels.
Headphones that audibly 'clip' mechanically (compress is more exact wording) I know that are quite widely known are the Koss thingies (KSC35,75,PortaPro) which is clearly seen in graphs.
Of course, as Amir also often mentions, is that you need quite a lot of output voltage when you want 'impressive' sound.
So your observations of having a lot of headroom (wouldn't call it power) and output voltage mimics yours an mine as well as Amir's.

However, the observation Amir made about the L30 is that its output voltage is enough to reach that. This isn't even a balanced amp and amps I use and design reach similar output levels, They will show more distortion and noise on the test bench but I don't hear any noise and distortion products are low enough to be masked.

From the measurements I have made (including compression tests which is stretching my limit) I can see all headphones react differently. Some are very good at this others are not. Mostly in the bass area. I don't perceive it that way even when reaching high levels shortly. Some headphones with high distortion in the area above 500Hz I can hear turning nasty or HP's like the Koss, K500/K501 and a Stax SR5.

You would need a test rig like the one Amir is testing to measure distortion at impressive levels. I know that reaching 110dB SPL peak in the bass is bound to create high levels of distortion in many headphones. All I can say is look into headphones that show low amounts of distortion measured at 100dB.

Stating that headphone amps aren't solved because we need even more power (you mean output voltage in case of the Senns) is incorrect though.
Most people enjoying music rarely will exceed a few mW. People that want to turn it up to impressive levels for short moments do need a LOT more power. Just not in > 5 Watts.

 

[Fluffy]

What you say makes sense. There's no reason for tiny drivers to need huge power. But it doesn't reflect when I listen. Like I said I've hit audible clipping. I can also hear - not sure the correct term -saturation. When things become compressed, highs become distortion and strongly sibilant. While my much more powerful amp doesn't exhibit that. It's a speaker amp Bas-x A-100 in direct drive mode getting the full power of the headphones. Though when I do add a dropping resistor to drop its power output (albeit drastically increase its output impedance to 220 ohms), I can also easily hit clipping/distortion.

What was the class A monoblock option you were talking about? I only know of one class A headphone amp and that's the unreleased Emotiva headphone amp and it is not a monoblock.

Needing the direct drive of a Bas-x A-100 for HD800/700 is pretty ridiculous. For the HD800s you'll get to 120db SPL at 170 mw, and for the HD700 you'll need 210 mw. The Bas-x A-100 can reach 1300 mw for the HD800 and 2600 mw for the HD700, while both of them has a maximum power rating of only 500 mw. So not only it will drive them to driver clipping, it will reach painful listening levels way before that.

Let's compare that with three balanced headphones amplifiers currently on the market. For 500$ you can get the Topping A90 that will output 320 mw for the HD800. For 300$ you can get the THX AAA 789 that will output 800 mw. For 200$ you can get the Schiit Magnius that will output 1000 mw.
So basically, these headphone amps can output more power to the HD800 than its driver can handle, and beyond what will be audibly painful to listen to. There is absolutely no need for a speaker amp.

As a side note, and I'm trying to say this as non-offensively as possible, I want to express genuine concern that you find yourself needing to listen at such high volumes. This could be a sign of hearing damage, but I'm no expert on the subject.

 

[Get a hearing test]

Its true. The only difference on higher cost amp is more knobs and features and POWERRRRRRRRR. Having an option to fry your headphones is a nice feeling every time you turn the knob, just dont sneeze. Dacs seem to have done the same. 200 dollar headphone. 100 dollar amp 100 dollar dac gets you most of the way there.
Scientists hate when they get the answer right the first time and have nothing else to go on. Eventually well figure out how to jack into the brain directly, then there will be truly no difference of opinion except in taste.

 

[garbulky]

My measurement setup is not suited for this. It is designed to measure at 90dB SPL and clips at 105dB.
When you reach clipping levels of amps you either are using wimpy amps or you listen at abnormally high levels.
Headphones that audibly 'clip' mechanically (compress is more exact wording) I know that are quite widely known are the Koss thingies (KSC35,75,PortaPro) which is clearly seen in graphs.
Of course, as Amir also often mentions, is that you need quite a lot of output voltage when you want 'impressive' sound.
So your observations of having a lot of headroom (wouldn't call it power) and output voltage mimics yours an mine as well as Amir's.

However, the observation Amir made about the L30 is that its output voltage is enough to reach that. This isn't even a balanced amp and amps I use and design reach similar output levels, They will show more distortion and noise on the test bench but I don't hear any noise and distortion products are low enough to be masked.

From the measurements I have made (including compression tests which is stretching my limit) I can see all headphones react differently. Some are very good at this others are not. Mostly in the bass area. I don't perceive it that way even when reaching high levels shortly. Some headphones with high distortion in the area above 500Hz I can hear turning nasty or HP's like the Koss, K500/K501 and a Stax SR5.

You would need a test rig like the one Amir is testing to measure distortion at impressive levels. I know that reaching 110dB SPL peak in the bass is bound to create high levels of distortion in many headphones. All I can say is look into headphones that show low amounts of distortion measured at 100dB.

Stating that headphone amps aren't solved because we need even more power (you mean output voltage in case of the Senns) is incorrect though.
Most people enjoying music rarely will exceed a few mW. People that want to turn it up to impressive levels for short moments do need a LOT more power. Just not in > 5 Watts.

That's a great response. I am well aware that your knowledge on this topic is far greater than mine. Thank you for sharing. What do you mean by differentiating power from voltage? What significance does that mean? Also what does your points regarding current vs voltage signify?

 

[garbulky]

Needing the direct drive of a Bas-x A-100 for HD800/700 is pretty ridiculous. For the HD800s you'll get to 120db SPL at 170 mw, and for the HD700 you'll need 210 mw. The Bas-x A-100 can reach 1300 mw for the HD800 and 2600 mw for the HD700, while both of them has a maximum power rating of only 500 mw. So not only it will drive them to driver clipping, it will reach painful listening levels way before that.

Let's compare that with three balanced headphones amplifiers currently on the market. For 500$ you can get the Topping A90 that will output 320 mw for the HD800. For 300$ you can get the THX AAA 789 that will output 800 mw. For 200$ you can get the Schiit Magnius that will output 1000 mw.
So basically, these headphone amps can output more power to the HD800 than its driver can handle, and beyond what will be audibly painful to listen to. There is absolutely no need for a speaker amp.

As a side note, and I'm trying to say this as non-offensively as possible, I want to express genuine concern that you find yourself needing to listen at such high volumes. This could be a sign of hearing damage, but I'm no expert on the subject.

Though I listen louder on headphones due to a lower sensation of the room, I have measured my average listening level in room (on speakers) using a calibrated SPL meter. I listen at about 68db. When i want to crank it up, I struggle to get in the mid 70's. I find 90+db listening annoying.
If you have access to the HD700 and HD600, I suggest trying it out on the A-100. Crank up the A-100 and you will encounter distortion. Granted siunce the A-100 is very powerful it is LOUD. Put it in direct drive mode and you have a different sound. Granted the differing impedance probably has a lot to do with it.
I've listened to the headphone amps/ouptut of the Macintosh tube preamp, Emotiva XSP-1, Emotiva XDA-2, DC-1, Asus Xonar Essence ST, Xenos 3HA, , Fire tablet, most iphone and ipad models below the X models.
All sound stressed or compressed at high volumes. You know that "dull sensation or rolled off treble" people complain about sennheisers? Well you get it on all of these. The headphones themselves feel a little closed in or dull on them. On several of these I have also been able to get audible clipping. Like...clipping that's not easily mistakable.

Out of them, The Emotiva DC-1 was probably the cleanest unit I've heard being low noise unfortunately it was not able to sound quite as open as the A-100 and did compress on louder sounds.
The three most powerful units I've heard were the Emotiva A-100 (top), the Mytek Liberty DAC (very clean sounding got just a hair harsh sounding compared to the A-100), the LH Labs Geek Pulse Infinity...which had tremendous power and no sense of compression...just didn't care for the overal dacs sound. Since it didn't have an analog input, I took it out of the running.
The A-100 sounds quite nice to me and is easily the best headphone amp I've heard but its weakness is an audible hiss on silence. I haven't had any problems driving it on direct drive mode. Perhaps a slight bit of treble harshness...very slight, but I'm hoping my next purchase will meet all my requirements. At this point, I'm not convinced that headphone amps need to max out at a couple of hundred millivolts at 300 ohms to be sufficient.
Theory is good until you are listening to them and hearing clipping or a compressed sound. At that point I have to side with less clipping and a more realistic sound.

 

[Vini darko]

Though I listen louder on headphones due to a lower sensation of the room, I have measured my average listening level in room using a calibrated SPL meter. I listen at about 68db. When i want to crank it up, I struggle to get in the mid 70's. I find 90+db listening annoying.
If you have access to the HD700 and HD600, I suggest trying it out on the A-100. Crank up the A-100 and you will encounter distortion. Granted siunce the A-100 is very powerful it is LOUD. Put it in direct drive mode and you have a different sound. Granted the differing impedance probably has a lot to do with it.
I've listened to the headphone amps/ouptut of the Macintosh tube preamp, Emotiva XSP-1, Emotiva XDA-2, DC-1, Asus Xonar Essence ST, Xenos 3HA, , Fire tablet, most iphone and ipad models below the X models.
All sound stressed or compressed at high volumes. You know that "dull sensation or rolled off treble" people complain about sennheisers? Well you get it on all of these. The headphones themselves feel a little closed in or dull on them. On several of these I have also been able to get audible clipping. Like...clipping that's not easily mistakable.

Out of them, The Emotiva DC-1 was probably the cleanest unit I've heard being low noise unfortunately it was not able to sound quite as open as the A-100 and did compress on louder sounds.
The three most powerful units I've heard were the Emotiva A-100 (top), the Mytek Liberty DAC (very clean sounding got just a hair harsh sounding compared to the A-100), the LH Labs Geek Pulse Infinity...which had tremendous power and no sense of compression...just didn't care for the overal dacs sound. Since it didn't have an analog input, I took it out of the running.
The A-100 sounds quite nice to me and is easily the best headphone amp I've heard but its weakness is an audible hiss on silence. I haven't had any problems driving it on direct drive mode. Perhaps a slight bit of treble harshness...very slight, but I'm hoping my next purchase will meet all my requirements. At this point, I'm not convinced that headphone amps need to max out at a couple of hundred millivolts at 300 ohms to be sufficient.
Theory is good until you are listening to them and hearing clipping or a compressed sound. At that point I have to side with less clipping and a more realistic sound.

Maybe go for broke and get a benchmark power amp and run your headphones off that

 

[garbulky]

Maybe go for broke and get a benchmark power amp and run your headphones off that

I don't want to mess around with converters and adapters. I want a purposebuilt unit. My A-100 used to be called the Min-x. It was a speaker only amp. People used its speaker outputs and used adapters to run their headphones off it. - usually the hifiman headphones. I waited till Emotiva came to their sense and simply put a headphone out jack on the A-100. Then I bought it
Now....I am really interested in hearing the benchmark amp for speakers. I have had a really nice experience with the flea size Emotiva PA-1 which uses an ICE module and sounds super fab to me - very close to the sound of my 70 pound XPA-1 monoblock amps. So I would be interested in how those benchmarks would do.

 

[Vini darko]

I don't want to mess around with converters and adapters. I want a purposebuilt unit. My A-100 used to be called the Min-x. It was a speaker only amp. People used its speaker outputs and used adapters to run their headphones off it. - usually the hifiman headphones. I waited till Emotiva came to their sense and simply put a headphone out jack on the A-100. Then I bought it
Now....I am really interested in hearing the benchmark amp for speakers. I have had a really nice experience with the flea size Emotiva PA-1 which uses an ICE module and sounds super fab to me - very close to the sound of my 70 pound XPA-1 monoblock amps. So I would be interested in how those benchmarks would do.

Yeah id love to lay my dirty mits on a pair of benchmark power amps too. Nothing wrong with adaptors as long as one is carful with usage procedures. Can understand wating a designed solution though. Not sure many companies would brave enough to do it though.

 

[solderdude]

What do you mean by differentiating power from voltage? What significance does that mean?

The significance is that amplifiers are voltage sources and not power sources. Well.. they are power sources in the sense that amplifiers have a voltage limit and a current limit.

It all has to do with voltage efficiency and impedance. I get where you are trying to make. I am approaching this from a technical P.O.V.

Both a 0.2W in 300 Ohm and 15W in 4 Ohm amplifier will go equally loud on a 300 Ohm headphone.
The 15W amp in the case of a 300 Ohm headphone will give the exact same output voltage (7.8V in this example).
Connected to both amps the drawn power will be 0.2W (26mA).
Assuming the headphone amplifier can provide 30mA and the 15W amp can deliver 2A in both cases the amps will not deliver more than 26mA.
Despite one amp being rated 0.23W max. (in 260 Ohm) and the other amp 15W in 4 Ohm.

So more wattage 15W compared to 0.23W makes absolutely no difference for the 300 Ohm headphone. The 15W one will not be louder nor supply more power to the headphone.
I assume that's where the usual 'need more power' myth comes from. In case of the HD600/HD700/HD800 you don't need more 'power' but you need a higher voltage as that is related to how loud it can go.

Put it in direct drive mode and you have a different sound. Granted the differing impedance probably has a lot to do with it.

It has everything to do with it in this case. These headphone sound different, not only in loudness but also tonality because of it.

Assuming you don't listen at more than 90dB SPL and assuming high DR recordings (110dB peaks) you would not be needing more than 10V for the HD700 assuming you do not EQ the bass.
This is 0.6W in 150 Ohm and 25W in 4 Ohm.

 

[garbulky]

The significance is that amplifiers are voltage sources and not power sources. Well.. they are power sources in the sense that amplifiers have a voltage limit and a current limit.

It all has to do with voltage efficiency and impedance. I get where you are trying to make. I am approaching this from a technical P.O.V.

Both a 0.2W in 300 Ohm and 15W in 4 Ohm amplifier will go equally loud on a 300 Ohm headphone.
The 15W amp in the case of a 300 Ohm headphone will give the exact same output voltage (7.8V in this example).
Connected to both amps the drawn power will be 0.2W (26mA).
Assuming the headphone amplifier can provide 30mA and the 15W amp can deliver 2A in both cases the amps will not deliver more than 26mA.
Despite one amp being rated 0.23W max. (in 260 Ohm) and the other amp 15W in 4 Ohm.

So more wattage 15W compared to 0.23W makes absolutely no difference for the 300 Ohm headphone. The 15W one will not be louder nor supply more power to the headphone.
I assume that's where the usual 'need more power' myth comes from. In case of the HD600/HD700/HD800 you don't need more 'power' but you need a higher voltage as that is related to how loud it can go.



It has everything to do with it in this case. These headphone sound different, not only in loudness but also tonality because of it.

Assuming you don't listen at more than 90dB SPL and assuming high DR recordings (110dB peaks) you would not be needing more than 10V for the HD700 assuming you do not EQ the bass.
This is 0.6W in 150 Ohm and 25W in 4 Ohm.

That's a really great explanation. I hadn't thought of it in this way. I think what you said will reveal itself in many applications for me. Thank you for sharing.

 

[KeithPhantom]

What exactly is the effect of crosstalk if it does not add anything to the music?

It is just a measure to know how much of the signal is leaking to an incorrect channel.